Which of me following is the BEST action to perform before applying patches to one of the hosts in a high availability cluster?

Before applying patches to a host in a high availability (HA) cluster, which often requires a reboot, it is essential to maintain service continuity. The best practice is to gracefully move all active workloads, such as virtual machines (VMs), from the host undergoing maintenance to another available host within the cluster. This process, known as a manual failover or live migration, ensures that the services provided by the VMs remain online and accessible to users. Once the host is free of active workloads, it can be safely patched and rebooted without causing a service interruption. This procedure is a fundamental aspect of managing and maintaining HA environments.

A. Disable the heartbeat network: Disabling the heartbeat would prevent the cluster from monitoring node health, potentially leading to a "split-brain" scenario and disrupting the entire HA mechanism.
B. Fallback cluster services: Fallback is the action of returning services to their original host after maintenance is complete and the host is verified as healthy, not before.
C. Set the cluster to active-active: Changing the fundamental architecture of the cluster is not a standard or necessary step for routine patching and does not address the need to vacate the specific host. ---

1. Microsoft Corporation. (2023). Cluster-Aware Updating overview. Microsoft Learn.

Reference: In the "Description of the CAU process" section

the documentation outlines the automated procedure for patching cluster nodes. Step 2 is to "Drain the roles from the node

" which means moving the clustered roles (such as VMs) to other nodes in the cluster before applying updates. This directly supports the concept of failing over workloads.

2. VMware

Inc. (2023). Place a Host in Maintenance Mode. VMware vSphere 8.0 Documentation.

Reference: In the section "Working with Maintenance Mode

" the documentation states: "To service a host

such as to install more memory or to apply patches

you must put the host in maintenance mode... vCenter Server migrates all powered on virtual machines from the host to other hosts in the cluster." This action of migrating VMs is a form of failover to prepare the host for maintenance.

3. Red Hat

Inc. (2023). Configuring and managing high availability clusters. Red Hat Enterprise Linux 8 Documentation.

Reference: Section 9.1

"Putting a cluster node into standby mode

" explains: "When you put a node into standby mode

the resources that are running on that node are moved to another node in the cluster... You can put a node into standby mode when you need to perform maintenance on that node." This describes the same principle of moving resources (failing over) before maintenance.